characterization of the molecular genetic variation in wild and ......populations for aquaculture. i...
TRANSCRIPT
-
Characterization of the molecular genetic variation in wild and farmed Nile tilapia Oreochromis
niloticus in Ghana for conservation and aquaculture development
Gifty Anane-Taabeah
Dissertation submitted to the faculty of the Virginia Polytechnic Institute and State University in
partial fulfillment of the requirements for the degree of
Doctor of Philosophy
In
Fisheries Sciences
(Department of Fish and Wildlife Conservation)
Emmanuel A. Frimpong, Chair
Eric M. Hallerman, Co-Chair
Jess W. Jones
Donald Orth
September 18, 2018
Blacksburg, VA
Keywords: Oreochromis niloticus; Oreochromis aureus; Oreochromis mossambicus;
Phylogenetic analysis; mitochondrial DNA; DNA microsatellites; West Africa; GIFT strain
-
Characterization of the molecular genetic variation in wild and farmed Nile tilapia Oreochromis
niloticus in Ghana for conservation and aquaculture development
Gifty Anane-Taabeah
ABSTRACT (ACADEMIC)
The Nile tilapia Oreochromis niloticus is native to Africa and middle East, and is an important
source of nutrition for many in sub-Saharan Africa. Understanding the genetic diversity within
and differentiation among wild populations can help identify O. niloticus populations that are
imperiled and require directed management, especially because of increasing threats to the
species’ long-term persistence in the wild, including habitat destruction, overfishing, climate
change, and hybridization with farmed populations. Knowledge of the genetic variation among
wild populations also can contribute to foundation and selection of genetically diverse
populations for aquaculture. I assessed the genetic variation among tilapia populations using fin-
clips collected between December 2014 and July 2017 from 14 farmed sources, mostly
originating from cage farms on the Volta Lake, and 13 wild sources from nine river basins in
Ghana. I also conducted a laboratory growth experiment in Ghana with two wild populations to
evaluate the tolerance of different genotypes to high temperatures, to inform their development
for aquaculture in West Africa. I found that pure O. niloticus populations persist in the wild but
some have been extensively introgressed with the closely related species, O. aureus, which has
not previously been documented in Ghana. Additionally, some wild populations appear to have
recently declined significantly in numbers, likely due to overfishing and habitat modification, the
latter primarily as a result of illegal alluvial mining ongoing in Ghana. Analysis of the farmed
populations revealed that at least two farms were growing the unapproved genetically improved
farmed tilapia (GIFT) and related strains, and that escaped individuals are admixed into some
wild populations. The results of my laboratory experiment showed that O. niloticus populations
occurring in northern Ghana already may be adapted to warmer temperatures and could be
developed and used purposefully in aquaculture, taking advantage of their adaptation. To protect
remnant pure O. niloticus populations in the wild, timely conservation decisions should be made
-
and implemented. Protecting wild O. niloticus populations also would ensure that pure
germplasms are available to develop aquaculture stocks from native populations.
-
ABSTRACT (PUBLIC)
The Nile tilapia Oreochromis niloticus is an important food source for many people in Africa.
However, many wild populations may be at risk of population decline and extinction because of
increasing human activities such as overfishing and farming of non-native strains. Understanding
the genetic differences among wild populations and comparing them with farmed strains can
inform protection of wild populations and also help develop aquaculture strains using native
populations as genetic resources. I assessed the genetic differences among tilapia populations
using fin-clips I collected between December 2014 and July 2017 from 14 farmed sources,
mostly originating from cage farms on the Volta Lake, and 13 wild sources from nine river
basins in Ghana. I also conducted a laboratory study with two wild populations to test their
tolerance to high water temperature. My research showed that pure O. niloticus populations still
occur in Ghanaian rivers, but some have reproduced widely with a similar species, O. aureus,
which is not known to occur in Ghanaian rivers. I also found that some wild populations may
have reduced population sizes because of overfishing or because their environments have been
impacted by illegal mining occurring in almost all Ghanaian rivers. My results indicated that at
least two farms were growing the genetically improved farmed tilapia (GIFT) and related
varieties, some of which have escaped the farms and mixed with wild populations. The results of
my laboratory experiment showed that O. niloticus populations occurring in northern Ghana may
be adapted to warmer water temperatures and could be selectively bred and used in aquaculture.
The information generated from my research should help in making timely conservation
decisions, which should help protect the remnant pure O. niloticus populations in the wild and
contribute to developing aquaculture responsibly.
-
v
ACKNOWLEDGEMENTS
I am sincerely grateful to my academic advisor and Chair of my research committee, Emmanuel
A. Frimpong for believing in me and giving me the opportunity to pursue a Ph.D. degree. He is a
wonderful teacher and a talented researcher, and I aspire daily to be like him. Without
Emmanuel, this dissertation would not have been completed. I also would like to acknowledge
another mentor, Eric M. Hallerman, the co-chair of my committee, who took me under his wings
and patiently guided me through the field of genetics to ensure the successful completion of my
studies. I am grateful for his quick reviews and comments on my manuscripts and the many
lunches to which he so generously treated his lab. I am deeply grateful to Jess Jones and Donald
Orth for serving on my committee, providing ideas and feedback, and reviewing manuscripts for
a successful execution of my dissertation research.
Funding for my research and my support at Virginia Tech was provided by the Virginia Tech
Graduate School; Interfaces of Global Change (Global Change Center); AquaFish Innovation
Lab through USAID; The Rufford Foundation (UK); Aquaculture Association of Ghana; Raanan
Fish Feed, Ghana; Consulting in Aquaculture Research and Training for Development
(CART4D); and the Burd Sheldon McGinnes Graduate Fellowship through the Department of
Fish and Wildlife Conservation, Virginia Tech.
I am grateful to the following individuals for their immense support in my field and laboratory
work in Ghana: Nathaniel Gyasi Adjei, Theophilus Ghunney, Abigail Ebachie Tarchie, Anthony
Aliebe, Raphael Nsiah-Gyambibi, Emmanuel Nyamekye, Isaac Nyame, Iris Fynn and Yaw B.
Ansah. Stephen Floyd, Tim Lane, and Vinnie Siegel, all formerly of Virginia Tech, assisted with
the genetic analysis of the tilapia samples from the 2014-2015 field season. Clay Ferguson of
Virginia Tech assisted me with DNA extraction in summer 2017, and I am very grateful.
I would like to thank the Fisheries Commission of Ghana for partnering in this work. Mr.
Emmanuel Aryee, Deputy Director at the Fisheries Commission, strongly recommended my
research, which was instrumental in securing the Rufford Foundation Grant. I would also like to
thank Mr. Francis Adjei Pilot, Aquaculture Centre Manager, and all his staff for their dedication
to my research, for providing tanks and ponds for my fish stock, and for providing the
Akosombo strain fingerlings for experimentation.
-
vi
To the many local fishermen, field guides and fish farmers who assisted with fish sampling, I
owe a huge debt of gratitude. Without their effort, it would have been impossible to collect all
the data from the whole of Ghana. I am grateful to the drivers of the Faculty of Renewable
Natural Resources, KNUST, Ghana for their willingness to travel with me and also work as field
assistants when required.
I am grateful to all the VT-FIW professors and staff, graduate students, and Frimpong and
Hallerman lab students for making my stay at Virginia Tech a memorable one. I am deeply
indebted to the Frimpong family (Emmanuel, Sophia, Rachel, and Mark), Sheila Harris, Miluska
and J. Murray Hyde, Amber and Jeff Robinson and others I may have forgotten, for sacrificing
their time and babysitting my son to enable me work in the laboratory or write my dissertation.
I thank the P.E.O ladies of Blacksburg, Virginia, for the friendship, the annual Christmas Goody
Bag, and invitations to many social events. I was blessed with a wonderful church family at
Gateway Baptist Church, and I want to thank everyone that made me feel at home all these years.
Special thanks to Jeff and Susan Racow for upholding me in their daily prayers. I am deeply
grateful to Lorraine Proska and Barry Huehn for treating me like family and providing a
memorable home-stay experience for my son and I.
To my family in Ghana, I thank you for supporting me in all my endeavors and praying
constantly for me. To my husband John and my son Jude, I sincerely thank you for laboring with
me through this journey. I believe I can achieve anything because you believe in me. I love you!
-
vii
Table of Contents
ABSTRACT (ACADEMIC) ......................................................................................................................... ii
ABSTRACT (PUBLIC) ............................................................................................................................... iv
ACKNOWLEDGEMENTS .......................................................................................................................... v
LIST OF FIGURES ..................................................................................................................................... ix
LIST OF TABLES ....................................................................................................................................... xi
Chapter 1: General Introduction ................................................................................................................... 1
References ..................................................................................................................................................... 3
Chapter 2: High genetic variation among wild and farmed Nile tilapia in Ghana evident in mitochondrial
DNA .............................................................................................................................................................. 6
Abstract ......................................................................................................................................................... 6
Introduction ................................................................................................................................................... 6
Methods ........................................................................................................................................................ 9
Fish sampling and DNA extraction ........................................................................................................... 9
Polymerase chain reaction and DNA sequencing ................................................................................... 10
Data analysis .......................................................................................................................................... 11
DNA sequences and polymorphism ..................................................................................................... 11
Phylogenetic analysis .......................................................................................................................... 12
Results ......................................................................................................................................................... 12
D-loop haplotypes ................................................................................................................................... 12
D-loop phylogenetic relationships .......................................................................................................... 13
ND1 haplotypes ....................................................................................................................................... 15
ND1 phylogenetic relationships .............................................................................................................. 15
CO1 relationships ................................................................................................................................... 16
Discussion ................................................................................................................................................... 17
Native populations of O. niloticus in Ghana ........................................................................................... 17
New records of O. aureus in Ghana? ..................................................................................................... 18
The presence of GIFT and related improved strains in Ghana .............................................................. 20
Conclusion and implications ....................................................................................................................... 22
References ................................................................................................................................................... 23
Chapter 3: Recent genetic bottlenecks within wild tilapia populations in Ghana and the risk of admixture
with non-native farmed strains .................................................................................................................... 44
Abstract ....................................................................................................................................................... 44
Introduction ................................................................................................................................................. 45
Methods ...................................................................................................................................................... 47
-
viii
Populations sampled ............................................................................................................................... 47
Polymerase chain reaction and genotyping ............................................................................................ 47
Data analysis .......................................................................................................................................... 48
Results ......................................................................................................................................................... 50
Genetic diversity in wild populations ...................................................................................................... 50
Genetic differentiation among wild populations ..................................................................................... 51
Genetic variation in farmed populations and admixture with native populations .................................. 53
Discussion ................................................................................................................................................... 55
Genetic diversity within wild O. niloticus populations: evidence of recent bottlenecks ......................... 55
Genetic differentiation among wild populations: barriers to dispersal and a unique dispersal
mechanism............................................................................................................................................... 56
Genetic variation in farmed populations and admixture with native populations .................................. 58
Conclusions and Implications ..................................................................................................................... 59
References ................................................................................................................................................... 60
Chapter 4: Phenotypic variation for tolerance of high temperature in O. niloticus from the Volta River
basin ............................................................................................................................................................ 93
Abstract ....................................................................................................................................................... 93
Introduction ................................................................................................................................................. 93
Methods ...................................................................................................................................................... 96
Experimental location ............................................................................................................................. 96
Experimental fish .................................................................................................................................... 96
Experimental design and conditions ....................................................................................................... 96
Statistical analysis .................................................................................................................................. 97
Results ......................................................................................................................................................... 97
Water quality measurements ................................................................................................................... 97
Effect of temperature treatment and strains on growth and survival ..................................................... 98
Discussion ................................................................................................................................................... 98
Conclusion and Implications ..................................................................................................................... 100
References ................................................................................................................................................. 100
Chapter 5. General Conclusions and Implications .................................................................................... 111
References ................................................................................................................................................. 114
-
ix
LIST OF FIGURES
Chapter 2
Figure 1. Sampling locations for wild Oreochromis spp. collected in eight rivers and one coastal
lagoon in Ghana from December 2014 through July 2017. Sites are indicated with red triangles.
………………………………………………………………………………………………….29
Figure 2. Phylogenetic tree constructed from 391-bp D-loop sequences for Oreochromis spp.
using Bayesian analysis. The analysis was performed with 2 million Markov chain Monte Carlo
generations with four chains, a sample frequency of every 100 generations, and a burn-in of
500,000 generations. The analysis was performed in two runs. A total of 30,002 trees were
sampled. The average standard deviation of split frequency was 0.0073 with a -Ln likelihood of -
3687.36. Bootstrap support values are indicated to the left of each node. Photo credits: O. aureus
(Froese and Pauly 2018); all other photos are from this study.…………………………………30
Figure 3. TCS network of 12 native Oreochromis spp. haplotypes from nine drainages in Ghana
inferred from 391 bp of sequence from the mitochondrial D-loop region. O. aureus sequences are
included as outgroups. Inferred mutations are indicated by hash
marks………………………………………………….………………………………………...31
Figure 4. Phylogenetic tree constructed with 675-bp ND1 gene sequences for Oreochromis spp.
using Bayesian analysis. The analysis was performed with 1 million Markov chain Monte Carlo
generations with four chains, a sample frequency of every 100 generations, and a burn-in of
250,000 generations. The analysis was performed in two runs. A total of 15,000 trees were
sampled. The average standard deviation of split frequency was 0.0071 with a -Ln likelihood of -
2113.78. Bootstrap support values are indicated to the left of each node. All photos are from this
study……………………………………………………………….……………………………32
Figure 5. TCS network of three farmed tilapia haplotypes from Ghana (white shading), and the
GIFT and related strains of O. niloticus (gray shading) observed among 315 bp of the
mitochondrial CO1 gene sequence. PAC1_Ghana (native O. niloticus), O. aureus and O.
mossambicus sequences are included as reference samples to show the close relations between
the native O. niloticus and O. aureus populations; and the dissimilarity between O. niloticus and
O. mossambicus, and the genetically improved strains………………………………………...33
Chapter 3
Figure 1. Sampling locations for wild Oreochromis spp. collected at 11 sites from eight rivers
and one coastal lagoon in Ghana from December 2014 through July 2017. Sampling sites are
indicated with red triangles….………………………………………………………………….66
-
x
Figure 2. The Evanno et al. (2005) method supporting K = 5 as the most likely number of
clusters for wild O. niloticus populations studied. All inputs were based on Structure analysis
with MCMC = 500,000………………………………………………………………………67
Figure 3. STRUCTURE results for mean likelihood values suggesting K = 10 clusters (MCMC =
500,000) among the O. niloticus populations studied.……………………………………….68
Figure 4. STRUCTURE results showing the proportion of each wild O. niloticus population’s
ancestry that was inferred to have come from each of K = 5 (top plot) and K = 10 (bottom plot)
clusters (MCMC = 500,000). Populations are on the x-axis and Q values are on the y-axis. AF =
Afram, WB = White Volta-Binaba, WK = White Volta- Kantu, OT = Oti, AN = Ankobra, TA =
Tano-Asuhyea, TE = Tano-Elubo, J = Juen, BV = Black Volta, and LV = Lower
Volta………………………………………………………………………………………….69
Figure 5. Phylogenetic tree constructed from microsatellite DNA data for Oreochromis niloticus
collected from 11 sites in Ghana between December 2014 and July 2017. The tree was
constructed using the unweighted pair-group method with arithmetic mean (UPGMA; Sneath
and Sokal 1973) with distance measure DA (Nei 1983; Takezaki and Nei 1996; 2008) corrected
for population size. Bootstrap replication was 100,000……………………………………...70
Figure 6. Relationship between genetic distance and geographic distance among wild
Oreochromis niloticus collected from 11 sites in Ghana from December 2014 and July 2017.
………………………………………………………………………………………………...71
Figure 7. Relationship between genetic distances and logarithm of geographic distances among
Oreochromis niloticus collected from 11 sites in Ghana from December 2014 and July 2017.
………………………………………………………………………………………………...72
Figure 8. STRUCTURE results showing the proportion of each farmed and reference tilapia
population’s ancestry that was inferred to come from each of K = 2 (top plot) and K = 6 (bottom
plot) clusters (MCMC = 500,000). Populations are on the x-axis and Q values are on the y-axis.
AD = ARDEC, LE = Lee’s Farm, VC = Volta Catch, AT = Akosombo Tilapia Farm, FF = Fujian
Farm, LV = Lower Volta River, and BV = Black Volta River……………………………....73
Chapter 4
Figure 1. Comparison of growth of tilapia Oreochromis niloticus among three populations,
(Afram and Binaba stocks, and Akosombo strain) and three treatment levels of temperature.
……………...……………………………………………………………………………….105
Figure 2. Comparison of growth of tilapia Oreochromis niloticus at three treatment levels of
temperature for all populations (Afram and Binaba stocks, and Akosombo strain) combined.
………………………………………………………………………………………………106
-
xi
Figure 3. Comparison of growth of among three populations (Afram and Binaba stocks, and
Akosombo strain) of Nile tilapia Oreochromis niloticus.…………………………………….107
Chapter 5
Figure 1. Oreochromis spp. collected from different sites in Ghana showing morphological
variations.…………………………………………………………………………………….115
LIST OF TABLES
Chapter 2
Table 1. Sample and site information for Oreochromis spp. collected in Ghana from December
2014 through July 2017……………………………………………………………………….34
Table 2. Oreochromis spp. reference samples used in the D-loop and ND1 phylogenetic
analysis.……………………………………………………………………………………….35
Table 3. Variable nucleotide sites for haplotypes at the mitochondrial D-loop region for
Ghanaian Oreochromis spp. Alignment gaps are indicated by “-”. Nucleotides identical to those
in the first sequence are indicated with a dot…………………………………………………36
Table 4. Frequency distribution of D-loop haplotypes for all geographic sites, and the
corresponding species clusters inferred for Ghanaian Oreochromis spp.……………….........37
Table 5. Pairwise nucleotide uncorrected p-distances between haplotypes1 at the mitochondrial
D-loop region for Ghanaian Oreochromis spp. Representative samples for the haplotypes are in
parentheses…………………………………………………………………………………....38
Table 6. Variable nucleotide sites for haplotypes at the mitochondrial ND1 gene for Ghanaian
Oreochromis spp. Alignment gaps are indicated by “-”. Nucleotides identical to those in the first
sequence are indicated with a dot.……………………………………………………………39
Table 7. Frequency distribution of ND1 haplotypes for all geographic sites, and the
corresponding species and inferred clusters for Ghanaian Oreochromis spp..……………….40
Table 8. Pairwise nucleotide distances between haplotypes at the mitochondrial ND1 gene for
Ghanaian Oreochromis spp..………………………………………………………………….41
Table S1. Variable nucleotide sites for Oreochromis spp. haplotypes at the mitochondrial D-loop
region and the reference sequences used for the phylogenetic analysis. Alignment gaps are
indicated by “-”. Nucleotides identical to those in the first sequence are indicated with a
dot…………………………………………………………………………………………….42
-
xii
Chapter 3
Table 1. Sample and site information of Oreochromis spp. collected in Ghana from December
2014 to July 2017……………………………………………………………………………….74
Table 2: Technical details for amplification of ten microsatellite loci for O. niloticus from nine
river basins and five farmed sites in Ghana.………...………………………………………….75
Table 3. Summary of genetic variation among eight microsatellites deoxyribonucleic acid (DNA)
loci examined in wild tilapia populations from 11 sites (nine river basins) collected in Ghana
from 2014 to 2017. N = number of individuals genotyped per locus, Ho = observed
heterozygosity, He = expected heterozygosity, A = number of observed alleles per locus, Range =
base-pair differences between the shortest and longest microsatellite alleles observed per locus,
M-ratio = approximate ratio of A and Range, and FIS = inbreeding coefficient. Values in bold are
significantly different from each other (P < 0.05).…...…………………………………………76
Table 4. Effective population size (Ne) estimates and 95% confidence intervals for wild O.
niloticus population samples collected in Ghana. ∞ = infinite.…………………………………80
Table 5. Summary statistics and p-values for bottleneck tests conducted for wild O. niloticus
populations. Tests with significant p-values (α = 0.05) are indicated in bold………………….81
Table 6. Pairwise FST values from nuclear microsatellite DNA sequences for wild O. niloticus
populations sampled from 11 sites (nine river basins) in Ghana from December 2014 to July
2017.…….………………………………………………………………………………………82
Table 7. AMOVA for eight nuclear DNA microsatellites loci in wild tilapia populations collected
from 11 sites in Ghana from December 2014 to July 2017.……………………………………83
Table 8. Genetic divergence statistics for each loci of a given sample size (n). Genic
differentiation (G) from Fisher’s exact test, Hendrick’s GST’, Jost’s D are compared with FST
values.….…………………………………………………………………………………….…84
Table 9. Pairwise FST values from nuclear microsatellite DNA sequences for farmed O. niloticus
populations sampled from 5 farm aquaculture facilities in Ghana in 2017. AD = ARDEC, LE =
Lee’s Farm, VC = Volta Catch, AT = Akosombo Tilapia Farm, and FF = Fujian Farm.
...……………………………………………………………………………………………….85
Table 10. AMOVA for eight nuclear DNA microsatellites loci in farmed tilapia populations
sampled in Ghana in 2017...……………………………………………………………………86
-
xiii
Table S1. Allele frequencies (%) across 8 nuclear microsatellite DNA loci for wild O. niloticus
populations sampled from 11 sites (9 river basins) in Ghana from 2014 to 2017. AF = Afram,
WB = White Volta-Binaba, WK = White Volta- Kantu, OT = Oti, AN = Ankobra, TA = Tano-
Asuhyea, TE = Tano-Elubo, JU = Juen, BV = Black Volta, and LV = Lower Volta. Private
alleles are indicated in Bold…………………………………………………………………….87
Table S2. STRUCTURE results with estimates for probable K clusters for O. niloticus from
Ghana. K = 10 as the most likely number of clusters is indicated in bold……………………...92
Chapter 4
Table 1. Summary of treatment levels, growth and survival parameters measured for O.
niloticus.……………………………………………………………………………………….108
Table 2. Effect size and significance of fixed effects of temperature level and population of Nile
tilapia Oreochromis niloticus. The dependent variable specific growth rate and Tank ID is the
random effect………………………………………………………………………………….109
Table 3. Effect size and significance of fixed effects in temperature level and population of
tilapia. The dependent variable is specific growth rate and Tank ID is the random
effect.………………………………………………………………………………………….110
-
1
Chapter 1: General Introduction
The Nile tilapia (Oreochromis niloticus) is one of the most cultured fish worldwide (FAO 2018).
In Ghana and in many parts of Africa, O. niloticus also contributes significantly to total
freshwater capture fisheries. The success of O. niloticus as a global aquaculture species can be
attributed, in part, to many years of extensive research (Watanabe et al. 1985; Binh et al. 1996;
Garduño-Lugo et al. 2003; Ridha 2010; Mugo-Bundi et al. 2015). However, despite the wealth of
information available on the species, we are still limited in our understanding of the genetics of
natural populations of O. niloticus. Agnèse et al. (1997) presented one of the first studies that
utilized allozymes and mitochondrial DNA to cluster natural populations of O. niloticus in Africa
into three groups: 1) West African populations, 2) Ethiopian Rift Valley populations, and 3) Nile
drainage and Kenyan Rift Valley populations. Later studies using novel molecular genetics tools
and a suite of analytic methods have shown that fine-scale analysis within individual drainage
basins is required to identify conservation units within the species (Hassanien and Gilbey 2005;
Hallerman and Hilsdorf 2014).
Knowledge of the genetic variability within the species is important not only for
monitoring the diversity within natural populations in order to conserve the species’ genetic
resources, but is also vital for advancing aquaculture, through the selection of appropriate
resource stocks. This knowledge is particularly relevant to Ghana given a widespread concern
that the Akosombo strain (the eighth to tenth generation of a local selective breeding program;
Dewedar 2013) may not be a fast-growing strain. Additionally, anecdotal reports suggested that
some farmers in Ghana were already growing the Genetically Improved Farmed Tilapia (GIFT)
strain (Dey et al. 2000) or crossing the Akosombo strain with wild strains (Ansah et al. 2014).
The absence of clear guidelines for obtaining fingerlings for aquaculture hampers aquaculture
development in Ghana and may pose a significant threat to natural diversity in wild populations
of O. niloticus, since fish escapes from farms are common (Safo 2007; Fisheries Commission
2012; Attipoe et al. 2013; Hatchery International 2014).
Climate change also poses a threat to natural populations of O. niloticus. Among the
impacts of climate change on aquatic systems in Africa and other parts of the world are the
increase in temperature and an associated decrease in dissolved oxygen levels (Ficke et al. 2007).
Many tilapia species occurring in already-warm conditions in Africa may be threatened by an
increase in regional temperatures (Case 2006). The climate in the north of Ghana and the
-
2
northern sections of the Volta River basin is hot and drought-prone, compared to a relatively
cooler south. The annual mean temperature for the north is 29oC, while that for the south is 26oC
(Ghana Meteorological Agency 2016). As expected, the average water temperatures in rivers
vary along the latitudinal gradient of Ghana (Frimpong et al. 2016). The Intergovernmental Panel
on Climate Change projects a mean global temperature increase of 1.4 – 5.8oC at the close of the
21st century, if greenhouse gas emissions remain unabated (IPCC 2001). Ghana may experience
as much as a 2oC increase in annual temperature over the next century (Hulme et al. 2001). At
this projected rate of temperature increase, it is reasonable to expect that southern Ghana will
experience temperature conditions that approximate the current temperature conditions in the
north. Warmer temperatures in the south could present stressful environmental conditions to the
outdoor culture systems predominantly used in Ghana since fish farming is concentrated in
southern Ghana. However, an increase in water temperature also presents an opportunity for
aquaculture development if O. niloticus populations occurring in the northern latitudes are
already adapted to high temperature conditions compared to the Akosombo strain of O. niloticus,
which is widely farmed in Ghana (Frimpong et al. 2016). Thus, proactive interventions will be
necessary to ensure that the O. niloticus genetic resources are not compromised, will be
appropriately used in resilient aquaculture systems, and that the species persists in the face of
climate change.
Population and quantitative genetics studies with direct aquaculture applications, which
recognize the potential negative effects of climate change, could help conserve the natural
variability in O. niloticus populations in Ghana and promote the aquaculture industry. Frimpong
et al. (2016) recently used a population genetics approach combined with experimental studies to
provide fresh insights into the relatively understudied subject of tilapia population genetics
within their native range. The goal of their study was to identify O. niloticus populations with
potential adaptation to predicted future stressful climatic conditions for southern Ghana. The
authors conducted water quality sampling in three rivers (Afram, Oti, and White Volta) within
the Volta River basin, collected fish samples for population genetic analysis, and investigated
phenotypic variation for tolerance of high temperature, low dissolved oxygen, and high salinity
in O. niloticus from the three rivers. The authors found considerable molecular genetic
differentiation among the three wild populations of O. niloticus studied, and their results
suggested that at least one northern population of O. niloticus may be adapted to the high
-
3
temperature and low dissolved oxygen (DO) levels predicted for southern Ghana (Frimpong et
al. 2016).
Building upon these results, I conducted this research to characterize the genetic
variability of O. niloticus within different drainage basins in Ghana and assessed potential
genetic impacts of aquaculture on wild populations of O. niloticus to provide the basis for
conserving the local adaptation of natural populations. I also sought to identify populations with
desirable traits, such as tolerance to high water temperature, that can be incorporated into future
selective breeding programs.
References
Agnèse, J-F., B. Adepo-Gourene, E. K. Abban, and Y. Fermon. 1997. Genetic differentiation
among natural populations of the Nile tilapia Oreochromis niloticus (Teleostei,
Cichlidae). Heredity 79:89–96.
Ansah, Y. B., E. A. Frimpong, and E. M. Hallerman. 2014. Genetically-improved tilapia strains
in Africa: potential benefits and negative impacts. Sustainability 6:3697–3721.
Attipoe, F. Y. K., S. K. Agyakwah, R. W. Ponzoni, H. L. Khaw, and E. K. Abban. 2013. Genetic
parameters and response to selection in the development of Akosombo strain of the Nile
tilapia (Oreochromis niloticus) in the Volta basin, Ghana. Page 81 in the Book of
Abstracts, ISTA10 (10th International Symposium on Tilapia in Aquaculture), Jerusalem,
Israel, October 6–10, 2013. 220 p.
Binh, C.T., C. K. Lin, and H. Demaine. 1996. Evaluation of low cost supplementary
diets for culture of Oreochromis niloticus (L.) in North Vietman (Part I) –
formulation of supplementary diets. Thailand Special Topics Research 3. Fifteenth
Annual Technical Report submitted to the Collaborative Research Support Program. 9p.
Case, M. 2006. Climate change impacts on East Africa. A review of the scientific literature.
WWF-World Wide Fund for Nature, Switzerland.
Dewedar, R. 2013. Fast-growing fish variety could benefit Egypt and West Africa. The Science
and Development Network, accessed at http://www.scidev.net/en/agriculture-and-
environment/fisheries/ news/fast-growing-fish-variety-could-benefit-egypt-and-west-
africa-.html.
Dey, M. M., A. E. Eknath, L. Sifa, M. G. Hussain, T. M. Thien, N. Van Hao, S. Aypa, N.
http://www.scidev.net/en/agriculture-and-%20environment/fisheries/%20news/fast-growing-fish-variety-could-benefit-egypt-and-west-africa-.htmlhttp://www.scidev.net/en/agriculture-and-%20environment/fisheries/%20news/fast-growing-fish-variety-could-benefit-egypt-and-west-africa-.htmlhttp://www.scidev.net/en/agriculture-and-%20environment/fisheries/%20news/fast-growing-fish-variety-could-benefit-egypt-and-west-africa-.html
-
4
Pongthana. 2000. Performance and nature of genetically improved farmed tilapia: a
bioeconomic analysis. Aquaculture Economics and Management 4:83–106.
Ficke, A. D., C. A. Myrick, and L. J. Hansen. 2007. Potential impacts of global climate change
on freshwater fisheries. Reviews in Fish Biology and Fisheries 17:581–613.
Fisheries Commission. 2012. Ghana National Aquaculture Development Plan (GNADP).
Fisheries Commission, Ministry of Food and Agriculture. 85p.
Food and Agriculture Organization (FAO). 2018. Fisheries and Aquaculture Information and
Statistics Branch. Fisheries Global Information System (FIGIS). FI Institutional
Websites. In: FAO Fisheries and Aquaculture Department [online]. Rome. Accessed on
30 August 2018 at
http://www.fao.org/figis/servlet/SQServlet?file=/usr/local/tomcat/8.5.16/figis/webapps/fi
gis/temp/hqp_6915900564662940548.xml&outtype=html.
Food and Agriculture Organization (FAO). 2005. National Aquaculture Sector Overview. Ghana.
National Aquaculture Sector Overview Fact Sheets. Text by Awity, L. In: FAO Fisheries
and Aquaculture Department [online]. Rome. [Available at
www.fao.org/fishery/countrysector/naso_ghana/en).
Frimpong, E. A., S. Amisah, G. Anane-Taabeah, and A. Ampofo-Yeboah. 2016. Identifying
local strains of Oreochromis niloticus that are adapted to future climate conditions.
Report submitted to United States Agency for International Development (USAID)
Aquaculture Innovation Lab, Oregon State University, Corvallis, Oregon. 32pp.
Garduño-Lugo, M., I. Granados-Alvarez, M. A. Olvera-Novoa, and G. Muñoz-Córdova.
2003. Comparison of growth, fillet yield and proximate composition between Stirling
Nile tilapia (wild type) (Oreochromis niloticus, Linnaeus) and red hybrid tilapia (Florida
red tilapia×Stirling red O. niloticus) males. Aquaculture Research 34: 1023–1028.
Ghana Meteorological Agency. 2016. Climatology. Available at:
http://www.meteo.gov.gh/website/index.php?option=com_content&view=article&id=62:
climatology&catid=40:features&Itemid=67.
Hallerman, E., and A. W. S. Hilsdorf. 2014. Conservation genetics of tilapias: Seeking to define
appropriate units for management. Israeli Journal of Aquaculture – Bamidgeh
http://www.siamb.org.il/article-1290-IJA-66-2014-1043.aspx.
Hassanien, H. A., and J. Gilbey. 2005. Genetic diversity and differentiation of Nile tilapia
http://www.fao.org/fishery/countrysector/naso_ghana/enhttp://www.meteo.gov.gh/website/index.php?option=com_content&view=article&id=62:climatology&catid=40:features&Itemid=67http://www.meteo.gov.gh/website/index.php?option=com_content&view=article&id=62:climatology&catid=40:features&Itemid=67http://www.siamb.org.il/article-1290-IJA-66-2014-1043.aspxhttp://www.siamb.org.il/article-1290-IJA-66-2014-1043.aspx
-
5
(Oreochromis niloticus) revealed by DNA microsatellites. Aquaculture Research 36:1450
–1457.
Hatchery International. 2014. Two Ghanaian fish hatcheries. Accessed on November 24, 2016 at
http://hatcheryinternational.com/Profiles/two-ghanaian-fish-hatcheries/.
Hulme, M., R. Doherty, T. Ngara, M. New, and D. Lister. 2001. African climate change: 1900–
2100. Climate Research 17:145–168.
Intergovernmental Panel on Climate Change (IPCC). 2001. Climate Change 2001. Synthesis
report. Cambridge University Press, Cambridge, UK.
Mugo-Bundi, J., E. Oyoo-Okoth, C. C. Ngugi, D. Manguya-Lusega, J. Rasowo, V. Chepkirui-
Boit, M. Opiyo, and J. Njiru. 2015. Utilization of Caridina nilotica (Roux) meal as a
protein ingredient in feeds for Nile tilapia (Oreochromis niloticus). Aquaculture Research
46:346–357.
Ridha, M. T. 2010. Spawning performance and seed production from hybridization between
Oreochromis spilurus and the GIFT strain of the Nile tilapia Oreochromis niloticus.
Aquaculture Research 41:E723–E729.
Safo, P. 2007. Freshwater fish seed resources in Ghana. Pages 257–266 in M. G. Bondad-
Reantaso, editor. Assessment of freshwater fish seed resources for sustainable
aquaculture. FAO Fisheries Technical Paper. No. 501. FAO, Rome. 628p.
Watanabe, W. O., C.-M. Kuo, and M.-C. Huang. 1985. Salinity tolerance of Nile tilapia fry
(Oreochromis niloticus), spawned and hatched at various salinities. Aquaculture 48:159–
176.
http://hatcheryinternational.com/Profiles/two-ghanaian-fish-hatcheries/
-
6
Chapter 2: High genetic variation among wild and farmed Nile tilapia in Ghana evident in
mitochondrial DNA
Abstract
In many African countries where the Nile tilapia Oreochromis niloticus is the main aquaculture
species, native strains are widely considered inferior to the Genetically Improved Farmed Tilapia
(GIFT) strain. The need for improved strains has led to a wide call to introduce the GIFT strain
into Africa. However, the infrastructure for preventing the possible escape of farmed populations
into wild populations and ecosystems is generally lacking, and there are limited baseline studies
to guide conservation planning for wild populations. The present study was conducted in order
to: 1) characterize the genetic background of Nile tilapia O. niloticus populations from different
river basins in Ghana to provide baseline information on the species’ genetic differentiation to
guide management and conservation decisions, and 2) assess selected stocks of O. niloticus
currently being farmed in Ghana. A phylogenetic analysis of mitochondrial D-loop region, and
ND1 gene, and COI gene DNA sequences from 184 tilapia samples collected from 14 farmed
sources, mostly from cage farms on the Volta Lake; and 13 wild sources from nine river basins
in Ghana, revealed three genetically distinct groups (with 100% bootstrap support) of
Oreochromis spp. in Ghana: 1) pure O. niloticus populations in all rivers sampled with private
haplotypes in the Black Volta River and the Afram River, 2) established O. niloticus populations
introgressed with O. aureus mitochondrial DNA in almost all rivers, introgressed populations
that have not been previously described, and 3) the GIFT (genetically improved farmed tilapia)
strain of O. niloticus on some farms, introgressed with O. mossambicus, some of which have
escaped into surrounding rivers. The results underscore the importance of complementing
morphological traits with genetic data to identify populations and species of conservation
concern. In order to protect pure populations of O. niloticus, it is important to prevent all
upstream movement of farmed tilapia, especially since wild populations are already threatened
by natural hybridization with O. aureus.
Introduction
The Nile tilapia (Oreochromis niloticus) is one of the most widely introduced species outside its
native range. Originally from Africa and the Middle East, the Nile tilapia has been introduced to
nearly all tropical and sub-tropical regions, primarily for aquaculture purposes (De Silva et al.
-
7
2004; Eknath and Hulata 2009). Many genetically improved tilapia strains have been developed
for aquaculture production including GIFT, GET-EXCEL, BEST, GMT, Chitralada, YY-male,
COLD, and Florida red (Ordoñez et al. 2017). The Genetically Improved Farmed Tilapia (GIFT)
strain of O. niloticus was founded using parental stocks from eight countries, including four
African countries (Egypt, Ghana, Kenya, and Senegal), and developed by 15 generations of
selective breeding. The subsequent distribution of the GIFT strain for commercial culture in Asia
revolutionized tilapia aquaculture in Asia and contributed to increased global tilapia production
(Eknath et al. 1993; Eknath et al. 1998; Dey 2000; Eknath et al. 2007; Ponzoni et al. 2011).
However, due to the risk of contamination of locally adapted native genetic stocks, the
WorldFish Center and other development partners responsible for the GIFT strain adopted a
policy that did not allow the dissemination of GIFT to African countries where the original
parental stocks were collected (Gupta and Acosta 2004).
In Ghana, like in many other sub-Saharan African countries where O. niloticus is the
main aquaculture species, locally available native strains are widely considered inferior to the
GIFT and related strains with respect to growth performance. Recognition of the need for
improved strains of tilapia in Africa, and to ensure that Africa benefits from the GIFT project
without the associated ecological and genetic risks of introducing the GIFT strain resulted in the
development of the Ghanaian Akosombo and the Egyptian Abbassa strains using the GIFT
selective breeding methodology (Gupta and Acosta 2004; Attipoe et al. 2013; Dewedar 2013;
Mireku et al. 2017). In recent years, however, many commercial farmers in Ghana have
expressed discontentment with the growth and survival rates of the Akosombo strain. As part of
the development and validation of the Akosombo strain, the GIFT was imported by the Ghanaian
government’s Aquaculture Research Development Center (ARDEC) in 2012 for experimentation
alongside the Akosombo strain (Ansah et al. 2014).
Ansah et al. (2014) analyzed the economic benefits and the long-term ecological risks of
introducing the GIFT strain to Africa and suggested that practical biosecurity measures be
implemented prior to any future GIFT introductions. However, the development and
implementation of biosecurity measures would prove effective only if countries properly defined
their conservation goals through a careful evaluation of the differentiation of populations
requiring protection from genetic introgression in specific geographic regions. In Ghana, there
are unconfirmed reports of farmers growing GIFT even though the strain has not been officially
-
8
approved for commercial farming. Hence, it is urgent to assess O. niloticus genetic resources in
Ghana to inform conservation of any genetically distinct populations remaining.
A systematic assessment of the genetic differentiation of O. niloticus populations in
Ghana would be important not only for developing conservation goals and protecting locally
adapted wild populations from introgression with introduced improved strains, but also could
help improve aquaculture locally. Regardless of whether the government plans to continue
improving the Akosombo strain or develop a new improved strain, an understanding of the
genetic differentiation among wild O. niloticus populations would benefit future development of
regionally appropriate stocks for aquaculture. In the Ghanaian context, research focusing on the
genetic variation among O. niloticus populations using both mitochondrial and nuclear DNA
markers have been limited to the Volta system (Agnèse et al. 1997; Falk et al. 2003; Falk and
Abban 2004; Attipoe et al. 2013; Mireku et al. 2017), and have not explored the possibility that
samples morphologically identified as O. niloticus could be introgressed or different species.
Hybridization and gene flow occur easily in tilapiine cichlids, many of which have
undergone a recent evolutionary radiation and show evidence of incomplete speciation
(Wohlfarth and Hulata 1981; Trewavas 1983; Agnèse et al. 1998; Nagl et al. 2001; Cnaani et al.
2003; Romana-Eguia et al. 2004; D’Amato et al. 2007; Ndiwa et al. 2014). Both natural and
anthropogenic factors have contributed to gene flow between the three main tilapia species
cultured globally; Nile tilapia Oreochromis niloticus, Blue tilapia Oreochromis aureus, and
Mozambique tilapia Oreochromis mossambicus. The natural range of O. aureus largely overlaps
that of O. niloticus, which is widely distributed across Africa and Middle East (Trewavas 1983;
Rognon and Guyomard 2003). In contrast, the distributions of both O. niloticus and O. aureus do
not naturally overlap with that of O. mossambicus, which is restricted to southern African in
basins including the Lower Zambesi, the Brak River, the Boesmans River, and the Limpopo
system (de Moor and Bruton 1988; Nagl et al. 2001; D’Amato et al. 2007).
O. niloticus is native to the Senegal, Gambia, Niger, Volta, Benoue, and Chad basins
within West and Central Africa (Rognon and Guyomard 2003; Teugels and Thys van den
Audenaerde 2003). O. niloticus also occur in the Nile basins including many East African rivers
and lakes (Trewavas 1983; Rognon and Guyomard 2003), and in coastal rivers of Israel
(Trewavas and Teugels 1991). Within Africa, sympatric populations of O. aureus and O.
niloticus occur in the Senegal River, middle Niger River, and the upper tributaries of the Benoue
-
9
and Chad basins (Rognon and Guyomard 2003). However, in the Middle East, O. aureus is
restricted to the Nile delta, the Jordan basin, and some rivers in Israel (Rognon and Guyomard
2003). There are apparently no records of O. aureus in the Volta basin (Rognon and Guyomard
2003), nor in East Africa.
Natural hybridization in sympatric populations of O. niloticus and O. aureus has been
documented in West Africa in the Senegal River (Rognon and Guyomard 2003); and the two
species appear genetically more closely related than West African and East African O. niloticus
populations (Rognon and Guyomard 2003). Natural hybridization also has been documented
between O. niloticus and O. mossambicus in southern Africa (D’Amato et al. 2007), where the
former was introduced through aquaculture operations. For aquaculture purposes, many different
production strains have been produced by hybridizing O. niloticus, O. aureus, and O.
mossambicus, including the red tilapias widely used in Asia, which were produced by
hybridizing male O. niloticus and female O. mossambicus (Romana-Eguia et al. 2004).
Given how easily tilapias hybridize, it is important to combine both morphological and
genetic data to identify genetically pure O. niloticus populations in order to inform their
conservation. Therefore, this study was conducted in order to: 1) characterize the genetic
background of Nile tilapia O. niloticus from different river basins in Ghana to provide baseline
information of the species’ genetic resources in order to contribute to management and
conservation, and 2) assess strains of O. niloticus currently being farmed in Ghana using samples
from selected farms and the government-approved Akosombo strain.
Methods
Fish sampling and DNA extraction
The main morphological trait used to distinguish O. niloticus from sympatric tilapia species is
the presence of dark vertical bands on caudal fins (Eccles 1992; Teugels and Thys van den
Audenaerde 2003). Other distinguishing traits include the lack of enlargement of jaws in mature
males (Froese and Pauly 2018), and a non-tessellated genital papilla in breeding males
(Trewavas 1983). In this study, fish were morphologically identified as O. niloticus based on the
presence dark vertical bands on caudal fins. Fin-clips from nearly 700 fish were collected
between December 2014 and July 2017. The samples included wild individuals from 13 sites,
which represented eight major rivers and one coastal lagoon in Ghana: Afram River, Oti River,
-
10
White Volta River, Black Volta River, Pra River, Tano River, Ankobra River, Lower Volta
River and Juen Lagoon (Figure 1).
Farmed O. niloticus samples were obtained from 14 sources including selected farms and
commercial hatcheries, as well as wholesale distribution points and local markets (Table 1).
Samples of the Akosombo strain of O. niloticus were obtained from the Aquaculture Research
and Development Centre (ARDEC) and the Pilot Aquaculture Centre (PAC) as reference
samples for evaluating the farmed populations. All fin-clips were stored in paper envelopes,
dried, and transported to Virginia Tech for laboratory analysis. Tilapia from two U.S. grocery
stores, which had originated from China and Ecuador (Table 2), were purchased in October 2017
as reference populations for the GIFT strains of O. niloticus reportedly grown illegally in Ghana.
Fish samples from the neighboring country, Cote d’Ivoire, were obtained as West African
reference samples (Table 2).
Total genomic DNA was extracted using the DNeasy extraction kit (Qiagen) according to
the manufacturer’s instructions. The DNA was quantified using a BioDrop spectrophotometer
and concentrations were standardized for use in the polymerase chain reaction (PCR). For this
study, 184 samples were selectively sequenced and analyzed. The study samples included 93
wild samples, 53 samples from seven farms and hatcheries, and 38 market samples originating
from cage farms on the Volta Lake.
Polymerase chain reaction and DNA sequencing
Three mitochondrial DNA markers - the displacement loop (D-loop) region, the first subunit of
the NADH dehydrogenase (ND1) gene, and the cytochrome c oxidase subunit 1 (CO1) gene -
were selected and amplified by polymerase chain reaction (PCR). The D-loop was chosen
because it is the most variable region in the mitochondrial genome and known to be a “hot spot”
for mutation. The ND1 gene was chosen as it is a relatively more conserved region compared to
the D-loop region. CO1 was chosen because it is used widely as a barcoding gene marker to help
identify different putative species. Primers (forward 5’-ACCCCTAGCTCCCAAAGCTA-3’ and
reverse 5’- CCTGAAGTAGGACCAGATG-3’) previously designed for O. niloticus (Nyingi et
al. 2009) were used for the D-loop. New primers were developed for ND1 with the Primer 3
program in GenBank using O. niloticus ND1 sequences available in GenBank (He et al. 2011).
Of the ten candidate primer pairs generated by Primer 3, three pairs (ND1Pr1, ND1Pr5, and
-
11
ND1Pr10) were evaluated. A multiplex of ND1Pr5 (forward 5’-
TGCAACTACGAAAAGGCCCA-3’ and reverse 5’- TAAGTGCTAGGGTGAGGGCT-3’) and
ND1Pr10 (forward 5’- TAATCCTTCCCGCCTGACCT-3’ and reverse 5’-
GCCTGTTAGTGCGATTGGGA-3’) amplified nearly 80% of the entire ND1 gene, which
included the region amplified by the primer ND1Pr1. Thus, only ND1Pr5 (forward and reverse)
and ND1Pr10 (forward and reverse) were utilized for this study. Universal fish primers (forward
F2 and VF2; reverse R2 and FR1d; Ivanova et al. 2007) were used for CO1 gene amplification.
The 22-µl PCR amplification reaction consisted of 50-100 ng of genomic DNA, 5 U/µl
Taq DNA polymerase, 5x PCR buffer, 25mM MgCl2, 2.5 mM dNTP mix, 5 µM bovine serum
albumin, and 5 µM of primers. The CO1 amplification reaction used 10 µM of primers. The
following thermal cycling conditions were used for D-loop and CO1: 94oC for 3 min; 35 cycles
of 94oC for 40 sec, 52oC for 30 sec, and 72oC for 1 min; a final extension of 72oC for 5 min; and
a 4oC hold. Thermal cycling conditions for ND1 were similar to those for the D-loop except that
the annealing temperature was 56oC and the final extension time was 10 min. PCR products were
visualized with agarose gel electrophoresis to confirm amplicon sizes prior to sequencing with an
ABI3730 automated DNA sequencer at the Virginia Tech Biocomplexity Institute (Blacksburg,
VA).
Data analysis
DNA sequences and polymorphism – The raw DNA sequences were assembled using the
program Geneious® 11.1.2 (Biomatters Ltd., Auckland, New Zealand). The consensus
sequences obtained then were aligned and edited with the program ClustalW (Larkin et al. 2007)
embedded in GeneStudio ™ v2.2. (http://genestudio.com/). Variable sites and parsimony
informative sites were determined using MEGA7 (Kumar et al. 2016). The numbers of haplotype
sequences were determined separately for each mtDNA marker with the program DnaSP 6.11.01
(Rozas et al. 2017). Using the Basic Local Alignment Search Tool (BLAST) feature in
GenBank, haplotype sequences for both DNA markers were queried, and highly homologous
sequences obtained from the BLAST searches were retrieved and used as reference sequences
for the phylogenetic analysis.
All sequences were realigned and the best substitution model of sequence evolution was
determined using the program MrModeltest 2.3 (Nylander 2008) implemented within PAUP 4.0
-
12
(Swafford 2002). The best substitution model selected using the Akaike Information Criterion
was the General Time Reversible model with gamma-shaped distribution (GTR+G) for D-loop,
the Hasegawa, Kishino, and Yano model with gamma-shaped distribution (HKY+G) for ND1,
and the GTR with a proportion of invariable sites with gamma-shaped distribution (GTR+I+G)
for CO1. Phylogenetic analysis based on Bayesian inference was conducted with the program
MrBayes 3.2.6 (Ronquist et al. 2012) using the parameters specified from MrModeltest. The
resulting phylogenetic trees were visualized with the program FigTree 1.4.2 (Rambaut 2014).
Evolutionary divergence between haplotypes was estimated using pairwise percentage sequence
divergence (p-distance) implemented in the program MEGA7 (Kumar et al. 2016).
Phylogenetic analysis – Phylogenetic analysis was conducted using both phylogenetic
trees (D-loop and ND1) and haplotype networks (D-loop and CO1). The phylogenetic tree for D-
loop was constructed using the Ghanaian O. niloticus haplotype sequences, and six groups of
reference samples (Table 2). The reference samples were: 1) U.S. grocery store tilapia
originating from China and Ecuador , 2) GenBank sequences of the GIFT strain of O. niloticus
and Egyptian, Philippino, and American strains of O. niloticus which were studied in China, 3)
GenBank sequences of Oreochromis aureus and O. mossambicus, as well as O. niloticus x O.
aureus and O. niloticus x O. mossambicus hybrids, 4) West African reference samples
represented by O. niloticus samples obtained from Cote d’Ivoire and sequenced as part of the
present study, and a GenBank sequence of O. niloticus sampled from Senegal, and 5) GenBank
sequences of select subspecies of O. niloticus sampled from Kenya, representing East African
samples. All GenBank accession numbers are presented in Table 2. The phylogenetic tree for
ND1 was constructed with Ghanaian haplotypes and only reference groups 1 to 3 described
above. Coptodon zillii was chosen as the outgroup for both mtDNA markers. The Coptodon zillii
D-loop sequence was obtained from an individual collected from the Volta River in Ghana
during the field sampling. The Coptodon zillii ND1 sequence with accession number KM658974
was obtained from GenBank.
Results
D-loop haplotypes
A partial sequence of 391 bp covering the hypervariable region (280 bp) and the first part of the
central conserved region (111 bp) of the mitochondrial D-loop region was analyzed. There were
-
13
99 variable sites and 83 parsimony-informative sites. Thirty-eight haplotypes were identified
among the 184 individuals sequenced. The haplotype data showed 77 variable sites (Table 3),
and 68 parsimony-informative sites. Of the 38 haplotypes identified, 15 haplotypes were shared
among populations, while 23 were private (Table 4). Two haplotypes (haplotypes 2 and 3) were
exhibited among more than 50% (N = 95) of all the individuals, with every geographic site
exhibiting at least one of those two haplotypes (Table 4). Additionally, haplotypes 2 and 3 were
highly frequent in samples from the Volta system (White Volta River, Black Volta River, Oti
River, and Lower Volta River), and were exhibited by eight of the 10 ARDEC individuals. Nine
of the ten Ankobra River individuals exhibited haplotype 2. The other haplotype observed in the
Ankobra River sample was haplotype 33, which was shared with individuals from ARDEC and
the Afram River (Table 4).
Four of the shared haplotypes were noteworthy. Haplotypes 1 and 15 were observed only
in samples collected from wild populations. Haplotype 1 was shared by the White Volta River,
Black Volta River, Tano River, and the Juen Lagoon, while haplotype 15 was shared by the Oti
River and Black Volta River populations. Haplotype 9 was shared by the Tano River and the
Black Volta at Talewona, and haplotype 11 was restricted to the Tano River population (Table
4). Nine of the 23 private haplotypes were observed in five collections from the wild; Afram
River (haplotypes 36 and 37, N = 6), Black Volta River (haplotypes 29, 30, 31, and 32; N = 16),
Pra River (haplotype 12; N = 9), Lower Volta River (haplotype 5, N = 11), and Juen Lagoon
(haplotype 14, N =5). Additionally, haplotype 38, which was private to the Adjei Farm at Elubo,
may be characteristic of the Tano River. The remaining 13 private haplotypes mostly were
observed in collections from aquaculture facilities on the Volta Lake, and dominated by samples
from Fujian Farm (Table 4).
D-loop phylogenetic relationships
The results of the phylogenetic analysis of mitochondrial D-loop sequences showed four
genetically distinct clusters with 100% bootstrap support (Figure 2), which also were distinct
from that of the outgroup species, C. zillii. The first cluster, the O. niloticus cluster, contained 17
haplotypes, which clustered with one sample from Cote d’Ivoire (Onilo_CD), and the GenBank
sequence Kpa11 originally sampled from the Volta basin in Ghana by Falk et al. (2003) as part
of a genetic study of black-chinned tilapia, Sarotherodon melanotheron. The O. niloticus cluster
-
14
contained haplotypes 2 and 3 (the haplotypes observed in the majority of both wild and farmed
individuals), as well as the two “wild” haplotypes, 1 and 15 (Figure 2). Haplotypes 3, 10, 15, 31,
and 32 clustered with GenBank-derived sequences of O. niloticus x O. aureus hybrids (Figure 2).
The second cluster, the inferred O. aureus cluster, contained 11 haplotypes from
individuals from the Black Volta, White Volta, and Pra rivers, and nearly all the farm and market
samples. The O. aureus cluster also contained the Senegal sample, one sample from the U.S.
grocery-store tilapia reference group (Onilo_WA1), and all three O. aureus sequences from
GenBank (Figure 2). Two of the haplotypes in the O. aureus cluster were private haplotypes
from wild sources, haplotype 12 from the Pra River and haplotype 30 from the Black Volta
River. To assess the observation of samples morphologically identified as O. niloticus appearing
distinctly as O. aureus, the haplotype network constructed based on the method of Templeton,
Crandall, and Sing (TCS; Templeton et al. 1992); using TCS network (Clement et al. 2002) with
haplotypes from both the O. niloticus cluster and the O. aureus cluster. The results supported the
monophyletic grouping of O. niloticus and O. aureus previously observed from the Dloop
analysis, and showed genetic distinctiveness, with about 141 mutation steps between the two
clusters (Figure 3).
The third cluster, the O. niloticus x O. mossambicus cluster, contained a GenBank
sequence of a O. niloticus x O. mossambicus hybrid, six haplotypes including four private
haplotypes from three sites – Fujian Farm (haplotypes 22 and 24), the Lower Volta River
(haplotype 5), and Juen Lagoon (haplotype 19); haplotype 5 shared by the Fujian Farm, Lower
Volta River, and Akosombo tilapia farm; and haplotype 33 shared by ARDEC, Ankobra River,
and Afram River. The O. niloticus x O. mossambicus cluster also contained all five East African
reference sequences; seven of the nine U.S. grocery tilapia reference samples and the three
GIFT-related strains (Egyptian strain, Filipino strain and the American strain); and two samples
from Noe, Cote d’Ivoire. Haplotype 19 observed in the Juen Lagoon sample was identical to that
in the Onilo_Egypt strain and Onilo_eduardianus from East Africa. The third cluster may be best
described as an inferred O. niloticus x O. mossambicus cluster because of the common practice
of hybridizing male O. niloticus and female O. mossambicus to produce red tilapias (Romana-
Eguia et al. 2004). This may also explain why the individuals in this cluster appear
morphologically as O. niloticus (for example, the East African reference samples) but contain O.
mossambicus mtDNA.
-
15
The fourth cluster, the O. mossambicus cluster, consisted of all five GenBank sequences
for O. mossambicus; four haplotypes (private and shared) from the Fujian Farm and the
Akosombo tilapia farm; one of the U.S. grocery store tilapia samples (KB2); and one of the Noe
samples from Cote d’Ivoire. The GenBank O. niloticus_GIFT strain sequence and another O.
niloticus sequence also grouped with the O. mossambicus cluster.
Pairwise nucleotide p-distances provided further support for the clustering observed. The
genetic distances were considerably larger between clusters than within clusters (Table 5). For
instance, the genetic distances (i.e., dissimilarity) between the O. niloticus cluster and the O.
aureus cluster, O. niloticus x O. mossambicus cluster, and the O. mossambicus cluster were
about 38%, 57% and 66-71%, respectively, compared to the largest within-group distance of
1.6% for the O. niloticus cluster (Table 5). However, divergence based on fixed nucleotide
differences at the variable sites showed that individuals in the O. niloticus cluster were
genetically more similar to individuals in the inferred O. aureus cluster; and genetically distinct
from individuals in the inferred O. mossambicus clusters (Table 3). The variable sites of the D-
loop haplotypes in comparison to all reference samples are provided in Table S1.
ND1 haplotypes
Sixty individuals, including the nine U.S. grocery-store reference samples, were sequenced at the
ND1 gene, which represented a subset of the samples analyzed at the D-loop region. Sixteen
sequences (mostly farmed, market and reference samples) were excluded in the analysis because
they were either too short or too divergent. A total of 44 sequences with length 675 bp were
analyzed. There were 72 variable sites, 65 parsimony information sites, and 19 haplotypes. The
number of variable sites were the same for both the full dataset and the haplotype dataset. Table
6 provides haplotype and variable sites information. The 19 haplotypes observed included seven
shared haplotypes and 12 private haplotypes, and the pattern of individual groupings into
haplotypes closely matched the groupings for the D-loop sequences. The relatively high
percentage of ND1 private haplotypes was expected due to the small sample sizes for the sites
included in the analysis.
ND1 phylogenetic relationships
The phylogenetic relationships observed for the mitochondrial ND1 haplotypes were congruent
with the D-loop results, except that the clustering was not as distinct for O. niloticus and O.
-
16
aureus as was observed for the D-loop tree. Two main clusters were observed. Cluster one
contained O. niloticus, O. aureus and hybrids of the two species as evidenced by individuals that
grouped with the respective clusters in the D-loop tree and with the GenBank reference
sequences (Figure 4). Cluster two included two sub-clusters. The first sub-cluster, the inferred O.
niloticus x O. mossambicus cluster, contained the U.S. grocery store tilapia reference samples,
the GIFT-related strains, and two ND1 haplotypes from the Lower Volta River and Fujian Farm,
all of which had clustered similarly in the D-loop tree (Figure 4). The second sub-cluster, the O.
mossambicus cluster, contained the GIFT strain and two GenBank sequences from O.
mossambicus. Table 7 provides the haplotype frequency distribution across sampling sites.
The pairwise genetic distances calculated for ND1 haplotypes showed that the O.
niloticus/O. aureus cluster and the O. niloticus x O. mossambicus cluster were distinct.
Haplotypes in the O. niloticus/O. aureus cluster were separated from the O. niloticus x O.
mossambicus cluster by a distance of 9.2-10.6%, while the largest intraspecific distance between
haplotypes in the O. niloticus/O. aureus cluster and in the O. niloticus x O. mossambicus were
1.7% and 0.4%, respectively (Table 8). Genetic divergence based on the fixed nucleotide
differences at the ND1 sequence variable sites was congruent with the findings from D-loop
sequence analysis and showed the observed genetic similarity between individuals in the O.
niloticus / inferred O. aureus cluster; and the dissimilarity between the O. niloticus cluster and
the inferred O. mossambicus cluster (Table 6).
CO1 relationships
I verified the presence of GIFT and related strains in Ghana by sequencing the CO1 gene of two
farmed samples (FFARM28 and FFARM29) previously identified as GIFT haplotypes from the
D-loop and ND1 analyses. The COI sequences were compared with CO1 sequences from GIFT
samples from GenBank (accession numbers KU565825 and KU565864) and related strains (e.g.
EXCEL, BEST, PNTO4) collected in the Philippines (Ordoñez et al. 2017). This verification
step was necessary because no information on the GenBank Onilo_GIFT (used for the D-loop
and ND1 gene phylogenetic analysis) linked the sequence to the GIFT project, and no other
appropriate GIFT sequences were found on GenBank for the D-loop and ND1 analysis. I also
sequenced one farmed sample identified as a native haplotype (PAC1) as a reference sample.
The relationships among the samples were visualized using a haplotype network and were
-
17
congruent with the results obtained for both D-loop and NDI analyses. The results showed that
the Fujian Farm samples were distinct from the native stock. The Philippines GIFT strain
sequences clustered with FFARM29, other improved strains, and the Egyptian and GIFT-related
strains used in the D-loop and ND1 analyses (Figure 5). Additionally, FFARM28 clustered with
GenBank O. mossambicus sequences, similar to findings from the D-loop and ND1 analyses
(Figure 5).
Discussion
Native populations of O. niloticus in Ghana
The results of the phylogenetic reconstruction presented here suggest the presence of distinct
natural populations of O. niloticus in Ghana (Dloop_cluster 1). These distinct populations appear
to have originated in the Volta River basin, with ancestral populations from mainly the Black
Volta, White Volta, and the Oti rivers (e.g., haplotypes 1, 2, 3, 29, and 31). Populations in the
Afram River and the Lower Volta River were later established from these source populations
(Figure 3). The Tano River also appears to be a source population for O. niloticus in Ghana,
judging from the number of haplotypes shared with the Black Volta, White Volta, and Tano
rivers; and the fact that those rivers are not currently connected (e.g., haplotype 9 is shared by the
Tano River and the Black Volta River). The fact that one sample of O. niloticus from Cote
D’Ivoire (Onilo_CD) also clustered with the Ghanaian haplotypes may be an indication that O.
niloticus from the Tano River shared ancestry with that from Cote D’Ivoire. Additionally, it
appears that O. niloticus independently colonized the Pra and the Ankobra rivers through the
Tano River, although the mechanism of colonization should be explored since all three rivers
share a geographically restricted tilapia species, Tilapia pra (Dunz and Schliewen 2000). The
results of the study suggested that the Akosombo strain (ARDEC) is widely used by farmers
across the country, even though the use of wild strain was detected at farms in Elubo, Western
Region. This inference was supported by the haplotype distribution among sites and the number
of haplotypes that the ARDEC strain shared both with the source drainages, as well as with the
farm and market samples.
-
18
New records of O. aureus in Ghana?
The clustering of Ghanaian O. niloticus haplotypes with GenBank sequences of O.
aureus with 100% bootstrap support and the presence of sympatric populations of O. niloticus x
O. aureus hybrids suggest that O. aureus occurred in Ghana but had not yet been documented.
These are two possible explanations for this observation: 1) O. aureus populations occurred in
Ghana, but have not yet been identified because the species’ geographic distribution previously
was thought to not extend to Ghana (Rognon and Guyomard 2003), and 2) samples
morphologically identified as O. niloticus carried mtDNA nearly identical to O. aureus. Given
that the samples collected phenotypically appeared as O. niloticus, the occurrence of
introgressive hybridization could account for why they are genetically similar to O. aureus.
Introgressive hybridization is the result of hybridization and repeated backcrossing, leading to
the mtDNA of one species completely replacing the mtDNA of the other (Rognon and
Guyomard (2003).
Hybridization between O. aureus and O. niloticus is known (Agnèse et al. 1997; Rognon
and Guyomard 2003; Bakhoum et al. 2009), and introgressive hybridization has been reported
for the two species in West African rivers. For instance, Rognon and Guyomard (2003) found
that introgressed O. niloticus individuals from Senegal showed mtDNA identical to O. aureus,
but were distinct based on morphological and nuclear DNA analysis. These results also may
explain why the GenBank O. niloticus sequence from Senegal (accession number EF016715)
grouped with the O. aureus cluster in this study. This finding is congruent with the high genetic
differentiation observed by Nyingi et al. (2009) when the same Senegal-derived mtDNA
sequence was analyzed together with O. niloticus sequences collected in Kenya. D’Amato et al.
(2007) also reported introgressive hybridization between O. niloticus and O. mossambicus,
where morphologically identified O. mossambicus contained O. niloticus mtDNA.
The occurrence of private haplotypes of “O. aureus” in the Black Volta River and the
White Volta River further supports the possible presence of O. aureus populations in Ghana. In
Africa, O. aureus naturally occurs in the Lower Nile (Egypt), Chad Basin (Chad), Benue (middle
and upper Niger River; Nigeria, Niger, Cameroon, and Mali) and Senegal River (Senegal; Froese
and Pauly 2018). Thus, it is likely that O. aureus naturally colonized the Volta River basins from
the upper Niger River (Mali – Burkina Faso – Ghana). Further studies involving a detailed
-
19
survey of rivers within the Volta River basin should help assess the presence of O. aureus in
Ghana.
Another important factor to note is that O. aureus and O. niloticus show 88.2% similarity
at the non-coding D-loop region and 91.6% at mitochondrial coding genes (He et al. 2011). The
high nucleotide similarities between the species especially at coding genes could explain why the
D-loop provided higher resolution for observing the genetic distinctiveness between the two
species compared to ND1 and other coding genes (Ordoñez et al. 2017). However, although a
high bootstrap support (100%) was obtained for the D-loop tree, the relatively small nucleotide
differences observed in the O. niloticus and inferred O. aureus D-loop DNA sequences was
congruent with the ND1 results, and suggests that the two “species” may be two different
lineages within one species.
Rognon and Guyomard (2003) discussed two hypotheses, which could explain the low
mtDNA sequence divergence recorded between O. niloticus and O. aureus, and affect the
assessment of whether the two warrant separate designation as valid species. The two hypotheses
are: 1) the conservation of ancestral polymorphism, or 2) differential introgressive hybridization
(Avise 1994; Rognon and Guyomard 2003). The arguments supporting the first hypothesis
suggest that the speciation of O. aureus occurred relatively recently from the West African O.
niloticus group. However, the coalescence times estimates of 200,000 and 300,000 years
between the two species, based on previously reported substitution rates of 3.6% my-1
(Donaldson and Wilson 1999); and 5.6% my-1 (Nagl et al. 2001) respectively do not agree with
protein data (Rognon and Guyomard 2003), fossil records, morphology or biogeographical
history (Trewavas 1983). On the contrary, available fossil records and protein data suggest that
O. niloticus and O. aureus diverged about 3.3 ± 1.5 million years ago during the Upper Pliocene
(Trewavas 1983; Rognon and Guyomard 2003). Biogeographical information suggests that the
O. aureus ancestral populations occupied the Jordan basin and the Nile delta and encountered O.
niloticus ancestors from the Nile basins prior to their independent colonization of West African
river basins around Pleistocene times (Trewavas 1983; Agnèse et al. 1997; Lowe-McConnell
1998).
Given this divergence time, the second hypothesis supports introgressive hybridization
between O. niloticus and O. aureus, and would best explain the low divergence between the two
sympatric West African populations compared to the significantly high divergence between West
-
20
African and East African O. niloticus populations (Rognon and Guyomard 2003). These authors
suggested that long-term unidirectional introgression from O. aureus into O. niloticus is likely
the case in West Africa because O. aureus is outnumbered by O. niloticus in the region
(Trewavas 1983; Lowe-McConnell 2000). Hence, hybrids are likely the products of mating
between male O. niloticus and female O. aureus because unidirectional introgression most
frequently involves females from the less-common species (Avise 1994; Wirtz 1999). Also,
given that crossings between male O. niloticus and female O. aureus tend to produce relatively
more female offspring (Wohlfarth and Hulata 1983), unidirectional introgressive hybridization
may become recurrent, producing more and more females appearing phenotypically as O.
niloticus with O. aureus mtDNA.
The above discussions clearly support two assertions: 1) O. niloticus and O. aureus are
two valid species, which have naturally hybridized extensively within West African river basins
where they exist in sympatry; and 2) the likelihood of introgressed O. niloticus populations
occurring within the Volta River basin in Ghana is very plausible, which would require further
studies to ascertain their possible distribution. The fact that O. niloticus x O. aureus hybrids were
detected on all farms samples suggests that the parental populations that ARDEC used to develop
the Akosombo strain included such hybrids, which went undetected. Thus, through the
nationwide dissemination of the Akosombo strain, these hybrids have become widely spread
across the country.
The presence of GIFT and related improved strains in Ghana
The clustering of Ghanaian haplotypes of O. niloticus with O. mossambicus x O. niloticus
hybrids, U.S. grocery-store tilapia from China and Ecuador (D-loop cluster 3, ND1 cluster 2), O.
mossambicus sequences, and GenBank sequence Onilo_GIFT (D-loop cluster 4) suggests a
recent introduction of GIFT or the related improved strains into Ghana. The clustering of
Onilo_GIFT with O. mossambicus further indicates that the original GIFT included founding O.
niloticus populations which were introgressed with O. mossambicus prior to their inclusion in the
GIFT project. The most likely O. mossambicus hybrid populatio